Table. 1 Multi PGx-dedicated bioinformatics tools’ discrepancies report for each core pharmacogene in current study and the total measurement for such conflicts in 100 WES data.

From: Development of an extensive workflow for comprehensive clinical pharmacogenomic profiling: lessons from a pilot study on 100 whole exome sequencing data

Genes

CYP2B6

CYP2C19

CYP2C9

CYP2D6

CYP3A5

DPYD

SLCO1B1

UGT1A1

VKORC1

Same result in all tools

70

76

98

8

23

91

43

63

100

3 vs. 1*

15

11

0

1

1

0

45

0

0

2 vs. 2

4

2

1

0

0

0

7

0

0

2 vs. 1 vs. 1

5

2

0

12

0

0

4

0

0

2 vs. 1 (One tool did not call any diplotype)

6

9

1

40

75

6

0

19

Not applicable

All different

0

0

0

39

1

3

1

18

Not applicable

Total conflicts without 3 vs. 1

15

13

2

91

76b

9

12

37

–

  1. *Only “3 vs. 1” was not checked for further evaluations. No matched phenotype was removed from the final report. Tools’ report files for the rest of the “vs.” situations are checked manually against PharmVAR and PharmGKB. Wrong or non-clear calls are interpreted as not accepted calls and removed. The overall concordance rate for all tools: 71% (including 3 vs. 1 scenario) see the main text for more details.
  2. aStargazer VCF only mode for calling stars in CYP2D6 as a highly structural polymorphic gene is not preferred.
  3. bCYP3A5 alleles are defined in a different way in Stargazer and PharmaKU.
  4. cTools in SLCO1B1 (less), UGT1A1, and VKORC1 use different allele nomenclature. Hence, the major discrepancies came from different allele names.
Back to article page